POS There are a number of ways that Alliance can benefit from the data it is gathering. The fact that is still orders periodically according to normal patterns is a complete waste of this wonderful data that they are gathering. First, their normal case scenario should not be the end point of how they order but the beginning. The median order quantity for a given period will clearly have some variability, which implies that Alliance can calculate a standard deviation of this data. The standard deviation information will be the basis for ordering going forward.

The demand differs according to a number of factors. The more of these factors that Alliance can identify for a given product, the better it will be able to manage its order quantities. For example, it knows that demand for some products is seasonal, and for others weather-related. This means that given a forecast Alliance can determine the best order quantity for a given region. An extreme example -- right now it is in the Atlantic hurricane season. If a hurricane was forecast for Miami, one would expect that sales of bottled water would skyrocket. The minute the hurricane forecast is identified, the company should be sending extra bottled water to its South Florida stores. There are many products and weather conditions that have a correlation, and Alliance has the data to explore those correlations.

Promotions were determined to be another tricky area for Alliance, yet the company has the ability to determine correlations with promotions. Alliance can even calculate cross-correlations, such as figuring out if...

When connections are established, that will help the order quantity. For example, if Kraft is pressuring Alliance to discount barbeque sauces, Alliance may know that sales of ribs spike when barbeque sauce goes on sale. The response would then be to increase the order size of ribs in order to avoid stockouts of that high-margin item when the barbeque sauce sale occurs. Correlations can be established any time they exist -- I would imagine Alliance stores in Kansas City and San Francisco would want extra beer and snacks, since those cities are in the World Series. That information could have been established based on trends from prior World Series, regardless of what cities were represented. This data is powerful, and Alliance can leverage that power by learning about the standard deviations and correlations for all the items in its product line.
b.

All of this information can help Alliance reduce costs and provide better service. The above examples highlight how Alliance can use this data to avoid stockouts, which is good for customer service. The data can also be used to determine the optimal economic order quantity for any given point, thereby reducing inventory (Agarwal & Holt, 2005). Reducing safety stock levels reduces inventory holding costs. So this data can be used to improver ordering efficiency, which both improves service and reduces costs simultaneously.

c.

Alliance can couple its point-of-sale system with a loyalty program to learn…